You stop bouncing between a generator and an editor. Kling O1 does both. Load up to 7 stills and go.
Free account · ⚡50 on signup · No card
● Shot on Kling O1
Two takes off Kling O1. Both from a single prompt, no reference stitching.
5sKling O1 · 5s
“Make a wooden dining chair in a sunlit kitchen”
5sKling O1 · 5s
“Make a coffee bottle on a sunlit marble counter”
A dancer spins in a mirrored studio, her skirt flares and the reflections track every turn, soft daylight, slow dolly in
Generated on OpenClips · kling-o1 · September 2026
A cyclist descends a mountain switchback, the camera follows from behind and leans into each turn with him, pine forest and morning haze
Generated on OpenClips · kling-o1 · September 2026
A woman in her late twenties talks straight into her phone camera in a bright kitchen, handheld selfie framing, natural window light, she smiles and gestures mid-sentence, casual creator energy, no on-screen captions
Generated on OpenClips · kling-o1 · September 2026
Cinematic product shot: a matte black wireless earbud case turns slowly on a dark reflective surface, a single warm rim light sweeps across it, macro lens, shallow depth of field, slow push-in, plain unbranded case
Generated on OpenClips · kling-o1 · September 2026
A street market at noon, a vendor hands fruit to a customer while a cyclist passes in the foreground and a dog trots across frame, static wide
Generated on OpenClips · kling-o1 · September 2026
● Release notes
Chapter 01 · The engine
Text-to-video, reference video and shot extension in one model.
Chapter 02 · The edit
Plain prompts edit the clip. No manual masking or keyframing.
Chapter 03 · The cast
Identity of characters, props and settings holds across the take.
Chapter 04 · The stack
Insert a subject and change the background in one pass.
● How to
Three steps from a reference pile to a finished shot. The page works logged out. Sign-in gates the render only.
Make a wooden dining chair in a sunlit kitchen
Make a coffee bottle on a sunlit marble counter
01 ·
≈ 1 min
Press a theme to load a prompt from the sheet. Or describe the edit in plain words.
02 ·
≈ 2 min
Up to 7 stills, or the clip you want changed. Name the role each one plays.
03 ·
≈ 2 min
Pick 720p or 1080p. Confirm the ⚡ quote. Render, download, remix.
● The sheet
O1 carries seven references and edits clips. Kling 3.0 finishes at 4K with sound.
| Spec | Kling O1 | Kling 3.0 |
|---|---|---|
| Status | Live in the composer | Live |
| Resolutions | 720p · 1080p | 720p · 1080p · 4K |
| Clip length | 3–10s | 3–15s |
| Audio | Keeps the source clip's audio | Optional native audio |
| Reference images | Up to 7 | Up to 1 |
| Video input | Feature or base clip | — |
| Attachment roles | First frame, first and last, reference | — |
| Aspect ratios | 16:9 · 9:16 · 1:1 | 16:9 · 9:16 · 1:1 |
| From | ⚡77 | ⚡77 |
● Spec sheet
Figures reflect OpenClips defaults, September 2026.
Compare every model that renders sound.Compare every model that takes several references.Compare every model that takes a clip.
● Quoted first
3s · kling-o1
77
$0.77 at pack rate
Reference tests and short cuts.
5s · kling-o1
129
$1.29 at pack rate
The working length for social delivery.
10s · kling-o1
343
$3.43 at pack rate
The long take for finish work.
Video is paid per render, never per month. Kling O1 starts at ⚡77 for a 3s 720p clip and reaches ⚡343 for a 10s 1080p render. Your ⚡50 signup grant covers a first short take, not a campaign. Packs start at $10/1,000⚡. The composer quotes the exact ⚡ for your settings before anything renders.
● The price grid
Read the cell you plan to render.
| Length | 720p | 1080p |
|---|---|---|
| 3s | ⚡77$0.77 | ⚡103$1.03 |
| 4s | ⚡103$1.03 | ⚡137$1.37 |
| 5s | ⚡129$1.29 | ⚡172$1.72 |
| 6s | ⚡154$1.54 | ⚡206$2.06 |
| 7s | ⚡180$1.80 | ⚡240$2.40 |
| 8s | ⚡206$2.06 | ⚡274$2.74 |
| 9s | ⚡231$2.31 | ⚡308$3.08 |
| 10s | ⚡257$2.57 | ⚡343$3.43 |
Dollars are the pack rate: $10/1,000⚡. Plans buy credits cheaper. The exact ⚡ price shows on the Generate button before you spend.
● OpenClips MCP
No console and no syntax. Your agent briefs Kling O1 through the OpenClips MCP server. Generate or edit, one thread.
https://mcp.openclips.aiStep 01
One sentence is the brief. Subject, motion, edit.
Step 02
Up to seven references, the length and the ⚡ quote.
Step 03
Download it, remix it, or ship it. Credits settle automatically.
● Version history
Newest first, straight off the roster.
video · Live on OpenClips
Next-gen motion control and consistency.
video · Live on OpenClips
Multi-reference omni model — motion transfer and clip editing.
● Questions
Kling O1 is Kuaishou's unified multimodal video model for making and editing clips. Kling AI unveiled Kling O1 on December 1, 2025. It handles text, video, image and subject inputs in one engine. On OpenClips you reach it through the same composer as every other model, with clips of 3 to 10 seconds at 720p or 1080p and up to 7 references per job.
Yes, you can try Kling free. A new OpenClips account gets ⚡50 on signup, with no card required. Video renders draw from that wallet, so the grant covers a short first take rather than endless clips. When it runs out, top up with credits. Packs start at $10/1,000⚡. Nothing renews unless you pick a plan yourself.
Kling O1 starts at ⚡77 for a 3s clip at 720p. Longer and sharper costs more: a 5s take at 720p is ⚡129, and a 10s render at 1080p is ⚡343. The composer shows the exact ⚡ before you render, so nothing surprises you. Credits sit in one wallet shared by every model on OpenClips.
Upload the clip, then describe the change in plain words. You can also swap a subject or restyle the shot in the same request. Kling O1 works from prompts rather than masks or timelines, so a note like take the crowd out of the background is the whole workflow. Renders land in about 2 minutes. Check it, reword, run again.
No. Kling O1 writes no audio of its own. It can carry the original track of a clip you feed it, and the composer offers that as a switch on the render. If you want sound written for the shot, switch to Kling 3.0 in the same composer, which generates native audio and renders up to 4K. Both share one wallet and one library.
Up to 7 images per render. The composer also names the role each attachment plays: first frame, first and last frame, or plain reference. A clip can go in too, as the feature or as the base. Kling O1 retains the identity of main characters, props and settings, which is what keeps a cast consistent across shots.
Kling O1 renders 16:9, 9:16 and 1:1 at either 720p or 1080p. That covers landscape for YouTube, vertical for Reels and TikTok, and square for feed posts. Clips run 3 to 10 seconds and you set the pacing. Choose the ratio in the composer before rendering, since price depends only on length and resolution.
Kling O1 is as safe as any hosted video tool, and on OpenClips your renders stay in a private library. You sign in with an email. Nothing posts to social by itself. Generations follow the provider's content rules, and you can delete any clip from your library. Downloads are yours to keep and cut into your own edits.
● The receipts
Every link goes to the vendor's own page.
Kling O1 folds text-to-video, reference video and shot extension into one model.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0Kling O1 edits clips from plain prompts, without manual masking or keyframing.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0Kling O1 retains the identity of main characters, props and settings.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0Insert a subject and change the background in one pass.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0Kling O1 is Kuaishou's unified multimodal video model for making and editing clips.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0Kling AI unveiled Kling O1 on December 1, 2025.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0It handles text, video, image and subject inputs in one engine.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0You can also swap a subject or restyle the shot in the same request.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0Kling O1 works from prompts rather than masks or timelines, so a note like take the crowd out of the background is the whole workflow.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0Kling O1 reads images, video clips, subjects and text as executable prompts.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0Group scenes hold together, because each subject is tracked on its own.
https://ir.kuaishou.com/news-releases/news-release-details/kling-o1-launches-worlds-first-unified-multimodal-video-model-0One prompt. One take. The price on the button before you press it.
Generate with Kling O1Free account. ⚡50 on signup. Pay per render.
Or roll with a sibling model
You have a clip that is almost right. A stranger walks through frame. The jacket is the wrong color. Kling O1 takes the clip plus a sentence and hands back the fixed shot. Kling O1 reads images, video clips, subjects and text as executable prompts, so the reference and the instruction go in the same box and you skip the round trip to an editor.
Load up to 7 images. Faces stay put across camera moves. Props keep their shape. Group scenes hold together, because each subject is tracked on its own. That is what makes the model useful for series work: same character, new shot, next day. The composer also names the role each attachment plays, so a still can be the first frame, one of a first-and-last pair, or a reference.
Clips run 3 to 10 seconds at 720p or 1080p, in 16:9, 9:16 or 1:1. A render takes about 2 minutes, so you can iterate over a coffee. O1 generates no audio of its own. It can carry an input clip's original track through instead, and if you want sound written for the shot, that is Kling 3.0 in the same composer.
Compare vendors side by side in the AI Video Generator. Run the same brief on two models and keep the better take. Full rates live on the pricing page, and every model on the models hub shares one wallet and one library.
● Where to go next
